linux-kernel - Re: [PATCH][next] vpda: Fix memory leaks of msg on error return paths

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20210125102136.6e7dye5ucoe5qiw2@steredhat>
Date:   Mon, 25 Jan 2021 11:21:36 +0100
From:   Stefano Garzarella <sgarzare@...hat.com>
To:     Colin King <colin.king@...onical.com>
Cc:     "Michael S . Tsirkin" <mst@...hat.com>,
        Jason Wang <jasowang@...hat.com>,
        Parav Pandit <parav@...dia.com>, Eli Cohen <elic@...dia.com>,
        virtualization@...ts.linux-foundation.org,
        kernel-janitors@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH][next] vpda: Fix memory leaks of msg on error return paths

On Fri, Jan 22, 2021 at 02:52:35PM +0000, Colin King wrote:
>From: Colin Ian King <colin.king@...onical.com>
>
>There are two error return paths that neglect to free the allocated
>object msg that lead to memory leaks. Fix this by adding an error
>exit path that frees msg.
>
>Addresses-Coverity: ("Resource leak")
>Fixes: 39502d042a70 ("vdpa: Enable user to query vdpa device info")
>Signed-off-by: Colin Ian King <colin.king@...onical.com>
>---
> drivers/vdpa/vdpa.c | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
>diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
>index 9700a0adcca0..eb1f5a514103 100644
>--- a/drivers/vdpa/vdpa.c
>+++ b/drivers/vdpa/vdpa.c
>@@ -540,13 +540,15 @@ static int vdpa_nl_cmd_dev_get_doit(struct sk_buff *skb, struct genl_info *info)
> 	if (!dev) {
> 		mutex_unlock(&vdpa_dev_mutex);
> 		NL_SET_ERR_MSG_MOD(info->extack, "device not found");
>-		return -ENODEV;
>+		err = -ENODEV;
>+		goto err;
> 	}
> 	vdev = container_of(dev, struct vdpa_device, dev);
> 	if (!vdev->mdev) {
> 		mutex_unlock(&vdpa_dev_mutex);
> 		put_device(dev);
>-		return -EINVAL;
>+		err = -EINVAL;
>+		goto err;
> 	}
> 	err = vdpa_dev_fill(vdev, msg, info->snd_portid, info->snd_seq, 0, info->extack);
> 	if (!err)
>@@ -554,6 +556,7 @@ static int vdpa_nl_cmd_dev_get_doit(struct sk_buff *skb, struct genl_info *info)
> 	put_device(dev);
> 	mutex_unlock(&vdpa_dev_mutex);
>
>+err:
> 	if (err)
> 		nlmsg_free(msg);
> 	return err;

The patch looks okay, but reviewing it I figure out that if 
genlmsg_reply() returns an error, it also frees the sk_buff passed, so 
IIUC calling nlmsg_free() when genlmsg_reply() fails should cause a 
double free.

Maybe we should do something like this (not tested):

diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index 9700a0adcca0..920afcb4aa75 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -538,24 +538,29 @@ static int vdpa_nl_cmd_dev_get_doit(struct sk_buff *skb, struct genl_info *info)
         mutex_lock(&vdpa_dev_mutex);
         dev = bus_find_device(&vdpa_bus, NULL, devname, vdpa_name_match);
         if (!dev) {
-               mutex_unlock(&vdpa_dev_mutex);
                 NL_SET_ERR_MSG_MOD(info->extack, "device not found");
-               return -ENODEV;
+               err= -ENODEV;
+               goto err_msg;
         }
         vdev = container_of(dev, struct vdpa_device, dev);
         if (!vdev->mdev) {
-               mutex_unlock(&vdpa_dev_mutex);
-               put_device(dev);
-               return -EINVAL;
+               err = -EINVAL;
+               goto err_dev;
         }
         err = vdpa_dev_fill(vdev, msg, info->snd_portid, info->snd_seq, 0, info->extack);
-       if (!err)
-               err = genlmsg_reply(msg, info);
+       if (err)
+               goto err_dev;
+
         put_device(dev);
         mutex_unlock(&vdpa_dev_mutex);
  
-       if (err)
-               nlmsg_free(msg);
+       return genlmsg_reply(msg, info);
+
+err_dev:
+       put_device(dev);
+err_msg:
+       mutex_unlock(&vdpa_dev_mutex);
+       nlmsg_free(msg);
         return err;
  }
  

Thanks,
Stefano